EP0975209B1 - Machine agricole de recolte commandee par robot - Google Patents
Machine agricole de recolte commandee par robot Download PDFInfo
- Publication number
- EP0975209B1 EP0975209B1 EP98918120A EP98918120A EP0975209B1 EP 0975209 B1 EP0975209 B1 EP 0975209B1 EP 98918120 A EP98918120 A EP 98918120A EP 98918120 A EP98918120 A EP 98918120A EP 0975209 B1 EP0975209 B1 EP 0975209B1
- Authority
- EP
- European Patent Office
- Prior art keywords
- module
- crop
- steering
- harvester
- line
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Lifetime
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01D—HARVESTING; MOWING
- A01D41/00—Combines, i.e. harvesters or mowers combined with threshing devices
- A01D41/12—Details of combines
- A01D41/127—Control or measuring arrangements specially adapted for combines
- A01D41/1278—Control or measuring arrangements specially adapted for combines for automatic steering
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01B—SOIL WORKING IN AGRICULTURE OR FORESTRY; PARTS, DETAILS, OR ACCESSORIES OF AGRICULTURAL MACHINES OR IMPLEMENTS, IN GENERAL
- A01B79/00—Methods for working soil
- A01B79/005—Precision agriculture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
Definitions
- the present invention relates to agricultural machines and more particularly to harvesting machines which, as part of the harvesting process, cut the crop or the plants on which the crop was grown.
- the invention provides an automatic steering control capable of locating the crop line between cut and uncut crop, steering the harvesting machine along the crop line as it traverses a field, sensing the end of a crop row and, upon reaching the end of the crop row, controlling the turning of the harvester to begin a new traverse of the field.
- EP-A-0 801 885 discloses an automatic steering control which processes video images to locate the crop line between cut and uncut crop and adjusts the travel direction to make the harvester follow this crop line. The document does not disclose how the end of a crop row can be established.
- US-A-4,077,488 equally discloses a system which locates the boundary between cut and uncut crop. However this system is capable of processing only one image line. It is not capable of establishing the presence of the end of a crop row in front of the agricultural machine.
- a robotic control system for an agricultural harvester including at least one video camera mounted on the harvester for viewing the changing scene in front of the harvester and producing images of the changing scene as the harvester moves along a planned path, and a module for scanning each line of an image to determine if a crop line boundary is indicated in the line, characterized in that said module comprises an end of row detector module for determining, from the determination of which lines indicate a crop line boundary and which lines do not, the most likely line imaging the end of a crop row.
- a method of harvesting crop material with a crop harvesting machine having a robotic control system comprising the steps of:
- the row detector module can be used in a robotic control system for an agricultural harvester, the robotic control system comprising an execution monitor module which is sequenced through a plurality of finite states by trigger messages from a field coverage planner module, a global trajectory tracker module, a controller module, an end of row detector module and a field coverage monitor module, the execution monitor module activating at least one of the field coverage planner module, the global trajectory tracker module, the end of row detector module, a crop line tracker module or a local trajectory tracker module in each finite state so as to generate steering signals for steering the harvester along a path specified by a field coverage plan to cut all crop in a field.
- the global trajectory tracker module may generate votes indicating the preferred direction the harvester should be steered so as to follow the path and the crop line tracker module generates steering votes on the direction the harvester should be steered to follow a crop line between cut and uncut crop, the system further comprising a continuously running steering arbiter module for developing steering commands from all the steering votes, and a continuously running controller responsive to the steering commands for developing signals to differentially drive front wheels of the harvester to steer it.
- the robotic control system includes at least one and preferably two alternately active video cameras mounted on the harvester for viewing the changing scene in front of the harvester and producing images thereof as the harvester moves along the planned path, a frame grabber for grabbing one image at a time from the active camera, and a video processing computer for analyzing each image pixel by pixel according to a color discriminant function to locate a crop line between cut and uncut crop, the color discriminant function being adaptively updated by computing, as the discriminant used in analyzing one image, the Fisher linear discriminant in
- the image Prior to analyzing each image to locate the crop line, the image may -be compensated for shadows.
- the robotic control system may include at least one and preferably two alternately active video cameras mounted on a harvester for viewing the changing scene in front of the harvester and producing images thereof as the harvester moves along the planned path, a frame grabber for grabbing one image at a time from the active camera, and a video processing computer for analyzing each image pixel by pixel and scan line by scan line to locate, for each scan line, a crop line point between cut and uncut crop if there is a crop line imaged in the scan line, and determining which scan line is most likely imaging the end of a crop row.
- the robotic control system may include at least one and preferably two alternately active video cameras mounted on a harvester for viewing the changing scene in front of the harvester and producing images thereof as the harvester moves along the planned path, a frame grabber for grabbing one image at a time from the active camera, and a video processing computer for analyzing each image pixel by pixel to determine the probability that a region of the image surrounding each pixel is imaging an obstacle in the path of the harvester.
- Fig. 1 illustrates an agricultural harvester 10 in the process of cutting a crop in a field having a boundary 12.
- the uncut portion of the crop is indicated by the shaded area 14.
- the edge or division between the cut crop and the uncut crop in front of the harvester is generally referred to as the crop line and is designated by the reference numeral 16.
- the entire area of the field is not planted with crop.
- a headland having a width D is left unplanted at each end of the field and provides an area in which the harvester may turn around.
- the division between the headland and that portion of the field planted with crop is generally referred to as the "end of row" and is designated by the reference numeral 18.
- the invention is not limited to use in rectangular fields and the crop area to be cut may be irregular in shape.
- the harvester 10 may take many forms (combine, forage harvester, etc.) but experimentation was carried out using a model 2550 SpeedrowerTM sold commercially by New Holland North America Inc.
- This harvester is similar to the one described in U.S. patent no. 5,327,709 and has a cutter bar (not shown) which may be raised and lowered and which is carried slightly above ground level at the front of the harvester when cutting crop.
- a reel 20 sweeps the cut crop material into conditioning rolls (not shown) for further conditioning, and the conditioned crop is then discharged from the rear of the harvester in a windrow.
- the commercially available harvester is retrofitted with left and right wheel encoders 28,30 (Fig. 2) on its front wheels 22,24.
- left and right wheel drive motors 32 and 34 (Fig. 2) and an electrically controlled hydrostatic transmission 35 are added for differentially and reversibly driving the front wheels 22,24.
- the harvester as thus modified, is steered by driving the front wheels at different relative speeds.
- a throttle control 36 was implemented as the combination of a small electric motor, electromagnetic brake and a spring loaded cable (not shown), but in a commercial embodiment it is anticipated that a throttle actuator of conventional design will be used.
- Implement or cutter bar control 38 comprises a conventional solenoid-actuated hydraulic valve (not shown) for raising or lowering the cutter bar.
- the conventional operator-actuated controls 40 are retained in the illustrated embodiment but, as discussed below, these controls may be eliminated in an embodiment which permits operation of the harvester without a human operator.
- Wheel encoders 28,30 are of conventional design and may be any type of encoder capable of producing on the order of 10,000 counts or pulses per wheel revolution.
- the output signals from the wheel encoders are applied to a Physical Control Computer (PCC) 50 which computes the vehicle speed from these signals.
- PCC Physical Control Computer
- the harvester 10 is provided with at least one video camera 42 (Fig. 2) mounted on an arm extending outwardly from one side of the harvester.
- the camera 42 is supported by an arm mounted near the top of the operator's cab 26 and extending to the left side of the harvester at a height of about 4m above the ground.
- the camera 42 is supported so as to be slightly inboard with respect to the left end of the harvester cutter bar and is aimed so as to capture an image of the landscape in front of the harvester.
- the cutter bar will slightly overlap the crop line. This insures that narrow strips of uncut crop material are not left on the field.
- a second camera 44 is mounted on the right side of the harvester and a dual,channel digitizer and frame grabber 46 is controlled to select the video output signal from the appropriate camera.
- Cameras 42 and 44 need not be video cameras but may be any form of device capable of detecting the crop height or spectral distribution as discussed by Ollis and Stentz in Proceedings of IEEE Conference on Robotics and Automation (ICRA 96), April 1996, pp.951-956.
- the cameras are RGB video cameras equipped with auto-iris lenses.
- CCD color cameras manufactured by Sony Corporation were used in developing the invention. The cameras are calibrated, using the method described by Tsai in the IEEE Journal of Robotics and Automation, Vol. RA-3, No. 4, August 1987, pp. 323-344, to convert image pixel coordinates into real world positions within the scenes viewed by the cameras.
- the video output signals from camera 42 and 44 are fed to the dual channel digitizer 46 which is controlled by software commands from a Task Management Computer (TMC) 52 to alternately select the output signal from one camera or the other.
- TMC Task Management Computer
- the digitizer grabs a frame of the signal and digitizes it for further processing by the Video Processing Computer 48.
- VPC 48 analyzes the signal as subsequently described to track a crop line, detect the end of a row, detect obstacles in the path of the harvester, and locate a crop line after the harvester has made a turn in a headland.
- a Global Positioning Satellite (GPS) receiver 54 and an Inertial Navigation System (INS) 56 are connected to the PCC 50 via serial RS-232 links 58 and 60, respectively.
- the computers 48, 50 and 52 are interconnected by a serial Ethernet link 62.
- PCC 50 and the TMC 52 may, for example, be Motorola MV162 (MC68040-based) computers running VxWorksTM.
- VPC 48 may be a Pentium-IITM based IBM-PC compatible computer running Microsoft Windows TM-NT.
- the GPS 54 receiver may be a NovAtel AG 20 GPS receiver.
- the INS 56 may be a fiber-optic gyroscope such as the AutoGyroTM sold commercially by Andrew Corporation.
- the hardware mentioned herein is by way of example only, it being understood that other computers, inertial sensors and GPS receivers may be used in practicing the invention.
- FIG. 2 Also shown in Fig. 2 are an optional base station computer 70 communicating with computers 48, 50 and 52 via a radio ethernet link , an optional operator display 72, and a safety monitor board 74.
- a harvester according to the invention may be completely robotically controlled and need not even have an operator's cab. All control is exercised from the computer 70 and the operator controls 40 and display 72 are not provided.
- the invention may be implemented in an on-board system which permits automated control of some operations and manual control of other operations by an operator. For example, an operator may manually steer the harvester to the beginning of a crop row and initiate robotic control of crop line tracking, the operator resuming manual control when the end of a row is reached.
- computer 70 is not required but the operator display 72 and operator controls are provided in the operator's cab.
- the display 72 is controllable to display various information such as an image of the scene viewed by a camera (Fig. 3) or various status information (Fig. 4) including cutter status (on/off and up/down), engine RPM, and harvester heading and orientation and operation mode.
- the display includes various soft keys (Fig. 3) operable to select various data for display, or to set various parameters.
- the requirements for the safety monitor 74 vary depending on whether or not provision is made for on-board operator control.
- the safety monitor detects abnormal status conditions and either stops the harvester or signals the operator when an abnormal condition is detected.
- the safety monitor per se, is not a part of the present invention and will not be further described.
- the Physical Control Computer 50 continuously monitors the output signals from the wheel encoders 28 and 30 as previously mentioned, and also continuously monitors data originating from the GPS 54 and INS 56. Based on information gathered from these sources, PCC 50 calculates and broadcasts to other modules up-to-date navigation estimates of global (absolute) position, local (relative) position, heading and ground speed.
- the global position is expressed in Cartesian coordinates relative to a system-wide pre-defined coordinate system origin. This coordinate system roughly corresponds to the area of the continent where the system is operating, i.e. its origin is reassigned only when the harvester is moved on the order of hundreds of miles.
- the global position estimate is computed by PCC 50 by updating the GPS information collected from GPS receiver 54 with wheel encoder data collected from wheel encoders 28 and 30 and. This allows more frequent and precise updates than would be available from the GPS alone.
- the global position estimate is updated solely from data obtained from the wheel encoders 28 and 30. Although the accuracy of the position estimate based solely on wheel encoder data degrades with time and harvester speed, normal performance is restored automatically when the PCC 50 detects that GPS data is again available.
- the local position is expressed in Cartesian coordinates relative to an arbitrary coordinate system established or fixed every time the entire control system is started up. This coordinate system can be thought of as the fields in which the harvester is presently operating.
- the local position estimate is derived from data obtained from the wheel encoders 28 and 30 and INS 56.
- the heading broadcast by PCC 50 is a heading expressed relative to the global coordinate system . This heading is derived from a combination of the information obtained from INS 56 and GPS 54.
- the Physical Control Computer 50 also processes steering commands received from the Task Management Computer 52 and develops signals for steering the harvester.
- a Steering Arbiter program module run by the Task Management Computer 52 develops the steering commands for steering the harvester.
- a steering command comprises an odd-numbered set of values, limited in magnitude and having values which are the reciprocal of the radius of curvature of the commanded turn. For example, there may be 201 steering commands with command #1 corresponding to a curvature of -0.1 (maximum left turn) and command #201 corresponding to a curvature of +0.1 (maximum right turn). Command #100 in this example would correspond to zero curvature thus steering the vehicle straight ahead.
- a steering command produced by Task Management computer 42 is transferred via serial link 62 to the Physical Control Computer 50 which resolves the steering command and generates analog current signals that are fed to electrically controlled hydraulic servo valves in transmission 35.
- the valves allow proportional amounts of hydraulic power (flow) to drive the wheel drive motors 32 and 34 at their respective desired speeds.
- the harvester ground speed or velocity is equal to the average value of the left and right wheel speeds while the curvature of its path is proportional to the difference in the wheel speeds. With respect to ground speed, the total hydraulic power available, and therefore the maximum ground speed attainable, is proportional to the harvester engine speed which is determined by the setting of a throttle control 36.
- the PCC 50 also issues signals for setting throttle control 36 and actuating implement controls 38.
- operator-actuated controls 40 are provided to permit an operator to manually control the harvester or over-ride automatic control
- signals from the operator-actuated controls are processed by PCC 50 to steer the harvester, set its speed, control the cutter bar, etc.
- the system of the present invention employs a behavior based approach to control the harvester 10.
- the system software defines a finite state machine which cycles the harvester through a series of states such as cutting or turning. Each state defines which program modules are active in the state and also defines the trigger events that allow the system to transition from the current state to another state. While in a given state, several executing modules or behaviors are activated or deactivated. Transitions between states are triggered by events signaled by a module. Trigger events may be signaled from all modules within the system.
- the program modules for controlling the harvester comprise an Execution Monitor (EM) module 100, a Crop Line Tracker (CLT) module 102, an End of Row Detector (ERD) module 104, a Local Trajectory Tracker (LTT) module 106, a Global Trajectory Tracker (GTT) module 108, an Obstacle Avoider (OA) module 110, a Steering Arbiter (SA) module 114, a Field Coverage Planner (FCP) module 116, a Field Coverage Monitor (FCM) module 118 and a Controller module 120.
- the CLT, EOR, and OA modules 102, 104 and 110 are all executed in the Video Processing Computer 48.
- the EM, LTT, GTT, SA, FCP and FCM modules 100, 106, 108, 114, 116, 118 are all run in the Task Management Computer 52.
- the Controller module 120 is run in the Physical Control Computer 50.
- the Steering Arbiter monitor module 114, Field Coverage Monitor module 118, Obstacle Avoider module 110 and Controller module 120 run continuously, that is, concurrently with any other of the program modules.
- the Execution Monitor 100 serves as the main control unit for the system and defines the states.
- the EM 100 sequences through a plurality of states and in each state activates one or more other program modules so as to cut the field according to a field coverage plan specified by the Field Coverage Planner 116.
- the FCP 116 is supplied with surveyed GPS points which define both the geometry of the field to be cut and the location of any known obstacles in the field.
- the planner is also supplied with data defining the dimensions of the harvester and its cutting implement, the maximum range of cameras 42 and 44, the harvester position in global coordinates, and the type of path to plan, for example a Path To Land plan, an S-type path, or some other type of path.
- Commands relating to points along the planned path are "tagged" with messages which result in the initiation of various operations as the harvester reaches each tagged point.
- a command may specify that the harvester is to move from point A to point B at a given turn radius of curvature and at some specified velocity.
- the next command may specify a turn to a particular heading, possibly a new velocity, and lowering of the cutter bar.
- the commands may also include such tag information as the optimal cutting width, or trigger messages which result in the activation of other modules by advancing the state of the Execution Monitor module 100.
- Fig. 6 illustrates the sequencing of Execution Monitor 100 between 8 states to move the harvester to a land section and cause the harvester to cut crop in the land section by traversing the land section in an S-type pattern, making a 180° turn at the end of each crop row.
- a land section may comprise an entire field or a section of a field separated from other sections by an irrigation border.
- State 1 represents the state of the system when the harvester is at some random position within a field and is planning a path to be traversed by the harvester to get to a certain land section within the field. It is assumed that the field has been surveyed to gather GPS data points which specify the geometry of the field, and these GPS data points have been processed by the FCP module 116 to divide the field up into land sections. In State 1 the EM module 100 instructs the FCP module 116 to plan a path to the next unprocessed land section within the field. Once FCP module 116 has constructed the path, herein referred to as the Path To Land or PTL plan, it signals a trigger message to EM module 100 which then advances to State 2.
- EM module 100 activates the GTT module 108 and instructs it to execute the PTL plan.
- the GTT module retrieves the PTL information from the FCP module 116 and begins executing the plan as later discussed in greater detail.
- the module signals a trigger message to EM module 100 which then advances to State 3.
- EM module 100 deactivates GTT module 108 and instructs the FCP module 116 to plan an S cut pattern for the given land section. Once the pattern has been planned, the FCP module first sends a message to the Crop Line Tracker module 102 informing the CLT module of the optimal cutting width to use when cutting the present land section. The FCP module 116 then sends a trigger message to EM module 100 to advance it to State 4.
- the EM module 100 activates both the GTT module 108 and the CLT module 102.
- the GTT module retrieves the S cut pattern to be tracked from the FCP module 116 and immediately begins tracking the path. That is, the GTT module starts generating steering votes (discussed later) which it sends to the Steering Arbiter module 114.
- the SA module 114 and the Controller module 120 are always active so the steering votes are converted, via modules 114 and 120, into signals which are fed to the transmission 35 (Fig. 2) to steer the harvester along the S path.
- the CLT module 102 When the CLT module 102 is activated, it first retrieves from FCP 116 an indication of whether to activate the left camera 42 or right camera 44 in order to track the crop line. The CLT module then begins processing images from the activated camera to find a crop line and to ultimately generate steering votes which are sent to SA module 114 to keep the harvester cutting crop with the appropriate cutting width sent to the CLT module during State 3.
- the EM module 100 also instructs Controller module 120 to prepare for cutting.
- the Controller module enables an autonomous mode allowing the software system to take over control of harvester motion, set the throttle up to a level appropriate for cutting, and turn on and lower the cutting implement. Once the controller has performed all of these functions it sends a trigger message to the EM module 100 to advance the EM module to State 5.
- the EM module 100 keeps the GTT and CLT modules active and both modules continue to produce the steering votes for steering the harvester.
- the EM module also instructs the Controller module 120 to slowly ramp up the speed to the optimal cutting speed for the present field and field conditions. Upon receipt of this instruction the Controller module starts ramping the speed of the harvester up to the requested speed. Upon reaching the requested speed, the controller holds this speed unless it is subsequently instructed to change speed.
- the Field Coverage Monitor 118 runs continuously. One of its purposes is to monitor progress of the harvester along the S path planned by the Field Coverage Planner 116.
- the FCM module 118 detects that the harvester is at, or has just passed, a tagged point on the path specifying that the harvester is near the end of the crop row, the module sends a trigger message to the EM module 100 to advance it to State 6.
- the EM module 100 keeps the GTT and CLT modules active and both of these modules continue to produce steering votes to keep the harvester on the S path.
- the EM module also activates the End of Row Detector module 104 which begins processing images from the active camera to detect the end of the row.
- the EM module is advanced from State 6 to State 7 only when it receives a trigger message telling it that the harvester is at the end of the row.
- This trigger message can from either the Field Coverage Monitor module 118 or the End of Row Detector Module 104. If the software determines that it has a good and valid global position the EM module 100 expects the end of row trigger message to come from FCM module 118 which will send the trigger message as a result of monitoring the progress along the specified path. Once the FCM module 118 detects that the harvester has reached the point along the path that is tagged the end of row point for that cutting pass, the FCM module sends the trigger message to EM module 100.
- the FCM module 118 can not trust the global position data and is therefore unable to determine when the harvester is at the end of the row.
- the EM module 100 expects the end of row trigger message to come from the End of Row Detector module 104 which sends the trigger message along with a measure of the distance from the point of detection of the end of row to the actual end of the row.
- the Execution Monitor module 100 deactivates the CLT and ERD modules 102 and 104, and instructs the Controller module 120 to lift the cutting implement. Then, based on whether or not the system has a valid global position, the EM module 100 either advances to State 8 immediately ( global position not valid) or waits to receive a trigger message from the FCM module 118 when the FCM module detects that the harvester is at the point tagged to be a spin turn point on the path. In State 8 the harvester performs a 180° spin turn about one front wheel to get the harvester aligned with the next crop row.
- this turn is controlled by either the Local Trajectory Tracker module 106 (global position not valid) or the Global Trajectory Tracker 108 (global position valid). This completes the cutting of one row and turning of the harvester at the end of the row.
- the module 106 or 108 that controlled the turn sends a trigger message to EM module 100 to transition the EM module back to State 4.
- the EM module is again sequenced through States 4-8 to cut the next row and again turn the harvester.
- the EM module is returned to State 1 to determine if there are any more land sections to be processed. If so, the FCP module constructs the Path To Land path necessary to guide the harvester to the new land section.
- the Field Coverage Planner 116 module plans a path for the harvester to follow.
- the path may be either a Path To Land path to move the harvester (1) from its current position to the end of the first row, (2) from the end of the last row of one land section to the first row of another land section, or (3) a coverage path such as an S-type path.
- the FCP requires, as input data, the type of plan it is to construct, surveyed GPS points which define both the geometry of the field and the location of known obstacles, including irrigation ditches within the field, the dimensions of the harvester and its cutting instrument, the maximum range of the cameras 42 and 44 and the current position of the harvester in global coordinates. From this data the FCP 116 constructs a path made of points which are tagged with position, velocity, curvature, trigger and other information, the points being with respect to the global reference frame.
- the FCP module When planning an S cut pattern for a land section, the FCP module computes the switchback path which requires the least number of passes while assuring that the land section is completely harvested.
- the FCP module computes the shortest path to get the harvester from its current position to the specified land section while avoiding obstacles and uncut crop. If a camera 42 or 44 detects an unexpected obstacle, for example a piece of machinery, or if the FCP module is notified by the Field Coverage Monitor module 118 that the harvester has deviated too far from the planned path, the FCP module plans a new path from the present position of the harvester to the original goal. FCP 116 merely plans a path, it does not execute the plan.
- the FCM module 118 functions as a safety monitor.
- the inputs to this module are the geometry of the field the harvester is operating in, the planned path from the FCP 116, and a maximum path tolerance value.
- the FCM module monitors the position of the harvester within the field at all times and if the harvester deviates from the planned path, or is found to be too close to an obstacle or an outer border of the field, the FCM module sends a halt command to the Controller module 120 to actuate the throttle control 36 to immediately stop the harvester.
- the FCM module Since the FCM module already monitors the harvester position with respect to the planned path for safety reasons, it is also used to check the planned path to determine when the harvester has arrived at a tagged point along the path. When the FCM module determines that the harvester is at a tagged trigger point along the planned path, it sends the trigger message associated with the tagged point to the EM module 100.
- the GTT 108 receives the current harvester position from Controller module 120 and the path to track from the Field Coverage Planner 116. From this information it generates steering commands which are sent to Steering Arbiter 114 to steer the harvester along the planned path and velocity commands which are sent to Controller module 120 for execution. If the path being tracked is a Path To Land plan, GTT 108 also sends point turn commands to Controller module 120.
- the LTT module 106 receives as inputs the current position of the harvester, in local coordinates, and the planned path. From this data it generates velocity commands which are applied to the Controller module 120 and steering commands or votes which are applied to the steering Arbiter module 114.
- the LTT module uses an adaptive pure pursuit tracking algorithm to generate steering commands allowing the harvester to accurately track the planned path. While tracking the path the LTT module monitors the tagged velocity points along the path planned. If the velocity of the harvester is found to be greater than what is specified in the plan the LTT module issues a command to Controller module 120 instructing it to set the velocity to the velocity value of the tagged path point the harvester is currently on.
- the purpose of the CLT module 102 is to analyze the video signal representing the scene, or a portion of a scene, within the field of view of the active camera 42 or 44 and generate an indication of where the crop line is relative to the center of the scene.
- the CLT module 102 comprises a discriminator and a segmenter, as described in the publication mentioned above, discrimination between cut and uncut crop being based on the ratio between two spectral bands (red/green) or the percentage intensity of green within a given spectral band.
- the discriminator computes from the digitized output signal of a camera a function d(i,j) of individual pixels within a window or region of interest of say about 400 pixels wide by about 480 pixels located at or near the center of the video camera imaging screen.
- Fig. 7A represents a plot of d(i,j) for a portion of one scan line near the center of the scan line, the discriminant being green percentage intensity. For each pixel position j the discriminator produces an output signal having a magnitude i related to the percentage green intensity at the pixel position. Because of the foliage on the uncut crop, the d(i,j) are clustered about a first mean value m r (Fig. 7B) for uncut crop and about a second mean value m l for cut crop. Fig.
- step function 7B suggests a step function defined by three parameters: j d , the j coordinate of the discontinuity; m l , the mean value of the step function to the left of the discontinuity; and m r , the mean value of the step function to the right of the discontinuity. Finding the best segmentation is thus a matter of finding the best fit step function (lowest least-squared error) to d(i,j) along a given scan line.
- the segmenter computes j d according to the equation: where e is the error.
- the discriminator portion of CLT 102 analyzes a pixel according to the discriminant function at step 121 and generates and saves a value i for that pixel.
- Step 122 determines if this is the last pixel to be examined in the current scan line and if it is not the last pixel, step 121 is executed to develop a discriminant value i for the next pixel. Assuming the window being analyzed is 400 pixels wide, steps 121 and 122 are repeated 400 times to develop and save 400 values i 1 -i 400 .
- the segmenter computes a best fit step function for these values. The location of the step defines a pixel location j n which in turn represents the location of the crop line as determined from the current scan line.
- the segmenter output signal is a binary signal having a bit position corresponding to each pixel on a single scan line and within the region of interest.
- the signal has in it a single 1-bit in the position corresponding to the computed step function pixel location j n .
- This bit represents a crop line tracker "vote" on the steering angle or direction in which the harvester should be steered so as to track the crop line.
- the votes on the steering direction should be in the same bit position of each output signal if the harvester is tracking the crop line.
- An accumulator is provided for each pixel position and at step 124 the vote generated at step 123 is added into the accumulator which accumulates votes for pixel location j n .
- Step 125 determines if the last scan line of the window has been processed. If the last scan line has not been processed, a return is made to step 121 and steps 121-124 are repeated to develop a vote on the steering direction for the next scan line. Step 125 detects when all 480 scan lines of the image within the window have been processed.
- the crop line tracker has developed 480 votes on the steering direction, these votes being distributed among 400 vote accumulators.
- the votes for groups of adjacent pixel positions are collected into 201 bins and at step 128 the bin vote counts are normalized.
- the CLT module produces 400 normalized votes on the steering direction, each normalized vote having a value between 0.0 and 1.0 and the sum of the normalized votes being 1.
- the normalized votes for pixel positions 1 to 200 that is the pixel positions to the left of the center pixel of the window, are assigned negative values so that these votes are votes for steering the harvester to the left of its present course.
- the normalized votes are supplied to the Steering Arbiter module 114 run in the Task Management Computer 52, and further processed by SA module 114 as described below to develop a steering command which is supplied to the PCC 50 where it is resolved to develop the analog signals controlling differential drive of the front wheels of the harvester.
- TMC 52 When TMC 52 receives a set of normalized votes from the CLT module 102 run in VPC 48, the TMC 52 sends a signal to the digitizer 46 so that the digitizer grabs another frame of the video signal from camera 42 or 44 for processing by the CLT module.
- the size of the window used in the crop line detector need not be 400 x 480 pixels and it is not necessary to process every pixel position of an image within a window.
- the window may be made smaller or, if the crop is relatively uniform over the field and is easy to discriminate from the cut portion of the field, the processing of every third or fourth pixel may be eliminated to reduce signal processing time in the CLT module.
- the CLT module 102 computes (step 126) the Fisher linear discriminant (see Pattern Classification and Scene Analysis, J. Wiley & Sons, 1973, pp. 114-118) in RGB space between the cut and uncut pixel classes. This discriminant is then used at step 121 in Fig. 8 as the discriminant for analyzing the next image.
- the Fisher discriminant computes a line in RGB space such that when the pixel values are projected onto that line, the ratio of average interclass distance to average interclass scatter is maximized. This results in the linear function which most cleanly separates the cut and uncut pixel classes.
- the discriminant function used in processing the first image is chosen arbitrarily and a poor choice may result in inaccurate crop line indications for the first few images as the algorithm converges to more effective discriminant functions.
- the CLT module 102 has a cycle time of about 5 Hz the crop line indications become quite reliable in about 0.5 second.
- the adaptive CLT alogrithm may be summarized as follows.
- Shadow noise can heavily distort both image intensity (luminance) and color (chrominance), resulting in errors in the crop line indications produced by the module. Shadow noise causes difficulties for a number of reasons. It is often not structured and thus is not well modeled by stochastic techniques. Its effects and severity are difficult to predict. For example, if the sun is momentarily obscured by a passing cloud or the orientation of the harvester changes rapidly, the prevalence and effect of shadow noise can vary dramatically on time scales of less than a minute.
- Normalizing for intensity will not solve the problem for two reasons.
- the primary reason is that it does not take into account the significant color changes present in the shadowed areas.
- the dominant illumination source for unshadowed areas is sunlight while the dominant illumination source for shadowed areas is skylight.
- a second reason is that it prevents the tracking algorithm from using natural intensity differences to discriminate between cut and uncut crop.
- the RGB pixel values at a given image point are a function of the spectral power distribution S( ⁇ ) emitted by a point in the environment.
- S( ⁇ ) spectral power distribution
- r o is a scaling factor
- r ( ⁇ ) is a function describing the response of the CCD chip and red filter of the camera, this function rapidly falling to zero outside a narrow wavelength band. Since r o and r ( ⁇ ) are purely functions of the camera it is merely necessary to construct a model of how shadows alter the function S( ⁇ ).
- the shadow compensation may therefore be implemented by (1) selection of appropriate constants for C red , C blue and C green , (2) determining whether points are shadowed or unshadowed, and (3) correcting the shadowed pixels using the above equation, that is, multiplying the shadowed pixel values by the respective correction factors. Intensity thresholding is used to determine whether a point is shadowed or unshadowed. Once an image has been compensated for shadows, it is then applied to the discriminator of Fig. 8.
- the crop line tracking algorithm shown in Fig. 8 segments images on the basis of color differences between the cut and uncut crop, the similarity in color between uncut crop and cut crop lying in a windrow produced by the harvester during the immediately preceding pass occasionally causes poor segmentation.
- This problem may be remedied by excluding from the image processed by the CLT module that region of the image where the windrow is located.
- the position of the windrow relative to the ideal crop line may be easily determined from the cut width on the previous pass and knowledge of the average width of a windrow.
- the cameras For the purpose of processing the relevant region of the video image, the cameras must be calibrated as mentioned above to capture the convergence of lines to the horizon as well as correct for camera lens distortion. Converting the real world distance between the crop line and the windrow into image values produces a cut-off line for both the segmenter and the learning phase of the adaptive algorithm. When the harvester is tracking using the left camera, all pixel positions j of the image to the left of this cut-off line are ignored when applying the discriminant function at step 121 of Fig.8, and when tracking with the right camera those pixel positions j of the image to the right of the cut-off line are ignored.
- the harvester may not be lined up perfectly with the crop when it completes a spin turn and this results in a poor crop line for the portion of the image which includes the already cut crop preceding the beginning of the crop.
- the initial imaged crop line is passed through a crop line assessor or "beginning of crop" detector to determine the position in the image where the uncut crop begins.
- the crop line assessor computes the regression line and error for each of four equal sections of the crop line.
- the top-most subsection of the crop line is most likely to be locked onto the crop hence its regression line is used as the base line for determining the skew of the lower subsections and it is never considered skewed.
- the beginning of the crop is considered to be at the end of the locked subsection that is lowest in the image.
- a Gabor elementary filter may be used as the texture operator.
- the Gabor elementary filter is a discrete realization of the function e (-(( x - x 0 )2+( y - y 0 ) 2 ))/2 ⁇ 2 ⁇ sin( ⁇ ( xcos ⁇ - ysin ⁇ )+ ⁇ where ⁇ , ⁇ , and ⁇ are the frequency, orientation and phase of the sinusoidal plane wave.
- the discrete realization forms a kernel that is convolved with the image. Kernels with orientations of 45 and 135 degrees are used.
- the resulting filtered image is then passed through the same segmentation process (step 123 of Fig. 8) applied to color discriminants.
- the operator controls 40 may include a switch or selection means to select the texture discriminant when sudan is to be harvested or select the color discriminant when another crop is to be harvested. In embodiments wherein no on-board operator controls are provided, the selection may be made via the off-board base station computer 70 (Fig. 1).
- the purpose of the Obstacle Avoider module 110 is to detect obstacles in the path of the harvester and, if an obstacle is detected, stop the harvester and signal the operator. Detection is carried out by analyzing, pixel by pixel, an image of the area in front of the harvester (which should be crop) to determine the probability that the region centered around the pixel belongs to a crop model.
- a discretized 2D histogram model of the crop is required for the analysis.
- the model is preferably acquired directly from the field to be cut by grabbing a frame image of the crop to be cut from camera 42 or 44 prior to starting the cutting of the field, the model being updated or trained with each new image.
- the step function fitting method described with respect to the CLT module 102 is used to generate the training data for building two probability density function (PDFs), one for the processed or cut portion of the field and one for the unprocessed or uncut portion of the field.
- PDFs are then combined into a single PDF for both the processed and unprocessed surface.
- Each pixel in an image is then examined and marked if it has a low probability of representing the processed or unprocessed surface.
- the Obstacle Avoider module 110 continuously repeats a sequence of steps 151-155 as long as no obstacle is detected.
- the obstacle avoider grabs a new image of the area in front of the harvester. That is, the output signal from a camera 42 or 44 representing one image is digitized.
- the OA module 110 selects (step 152) a pixel(i,j) that has not been processed.
- Step 153 using the histogram model and the pixel information for the pixels surrounding the selected pixel (i,j), computes the probability PROB(i.j) of the region of 10 x 10 pixels within the region centered around the selected pixel belonging to the crop model.
- PROB(i,j) is obtained by first multiplying together all points m,n within the 10 x 10 pixel region of i,j. PROB(i,j) is then obtained by dividing the number of points falling into histogram cell for pixel (m,n)'s color by the total number of points in the histogram.
- Step 154 compares the computed probability with a threshold value OBSTACLE_THRESH which is typically on the order of 10 to 200 and if the probability is not less than the threshold value it means that crop is being detected.
- Step 155 determines if the pixel just processed is the last pixel in the frame or image to be processed. If it is not, steps 152-155 are repeated. If step 155 determines that the last pixel of the image has been processed then the obstacle avoider returns to step 151 to obtain a new image for analysis.
- step 154 determines that PROB(i,j) falls below the threshold value OBSTACLE_THRESH then an obstacle has probably been detected. If not operating in the automatic guidance mode (step 156) the harvester is stopped by, for example, sending a signal to PCC 50 to control the harvester throttle control 36 (step 157). The operator may be notified by sending a command to PCC 50 to control the sounding of an alarm (not shown) and/or provide a visual warning. If the harvester is being operated in the automatic guidance mode then steering votes are generated (step 158) and supplied to the Steering Arbiter module 114. The steering votes generated at step 158 are similar to the steering votes produced by the crop line tracker in that there are 201 votes, one for each possible steering direction.
- one or more of the votes may be a veto vote having a value of -1.0.
- a veto vote on a given steering direction prevents the Steering Arbiter from generating a command to steer in that direction.
- the video cameras 42,44 are calibrated, the direction of each point in the camera field of view, relative to the current steering direction, is known from the pixel position where the point is imaged on the camera screen.
- the obstacle avoider computes what steering directions would cause the harvester to converge on the point imaged on that pixel.
- the obstacle avoider then inserts a veto vote in the bins corresponding to these steering directions when it generates the steering votes at step 158.
- the End of Row Detector module 104 detects when the harvester is near the end of a row and, when the GPS system is not functioning, provides a trigger message to the Execution Monitor module 100 indicating that the end of a crop row is near.
- the ERD module 104 also provides an indication of the distance between the current harvester position and the end of the row. When the end of row boundary is approximately perpendicular to the crop line and the cameras are mounted for zero roll, the distance to the end of the row is purely a function of the image row where the crop line boundary stops.
- the ERD module 104 is turned on by the EM module 100 when module 100 reaches state 6. When turned on, it resets a counter N (step 242 in Fig. 10), grabs an image of the area in front of the harvester (step 243) and removes shadow noise from the image as described above with respect to the Crop Line Tracker module. Module 104 then determines (step 244) the most likely end of row. The ERD module 104 determines the image row or scan line i which most likely separates those scan lines containing a crop line boundary, that is, a boundary between cut and uncut crop, from those scan lines which do not contain a crop line boundary.
- the ERD module analyzes each scan line according to the crop line best fit algorithm described with reference to the Crop Line Tracker module 102.
- the location and height of the resulting best fit step functions are compared to precomputed ranges stored in a memory in VPC 48 and derived from training data. If they fall within the allowed ranges, the boundary or step is accepted as a genuine indication of a crop line and the scan line is assigned a "beforeEnd” label. If no crop line is found, the scan line is assigned an "afterEnd” label.
- the scan line with the highest score S(i) is the scan line which is imaging the end of the row. From the camera calibration data, the distance between the current harvester position and the end of the row may then be determined.
- the computed distance is compared (step 245) with a stored distance value to determine if the end of row is eminent. If the computed distance is greater, the module does not send a triggering message to EM 100 but saves the computed distance and loops back to step 243 to grab and process a new image. Steps 243-245 are repeated until, at step 245, the computed distance is equal to or less than the stored distance value. When this occurs the computed distance for the present image is compared (step 246) with the distance value computed for the preceding image.
- the ERD module does not send a triggering message until it processes a plurality of images and finds that for N (say 3) images the computed distance for each succeeding image is less than the computed distance for the preceding image.
- step 246 determines that the computed distance for the present image is not less than that computed for the previous image, the program returns to step 243 to process another image. On the other hand if step 246 determines that the computed distance is less than that computed for the preceding image, counter N is incremented (step 247) and compared (step 248) with the value END_THRES (assumed to have a value of 3). If N is less than or equal to 3, the program loops back to step 243, but if it is greater than 3, a trigger message is sent (step 249) to the EM module 100, this message having in it the distance to end of row value computed for the last image processed.
- the reliability of the subsystem described above may be improved by providing a second subsytem in the form of a crop line assessor similar to the beginning of crop detector provided in the Crop Line Tracker module 102.
- the lines are evaluated from the bottom to the top with the bottom subsection always being considered as non-skewed. A possible end-of-row is considered to occur at the starting point (bottom of line) of the lowest skewed line.
- the possible end-of-row position outputted by the first subsystem is correlated with that of the second and a triggering message is sent to the EM module 100 only when the two outputs agree (and GPS data is currently unreliable or unavailable.
- Steering Arbiter module 114 One purpose of Steering Arbiter module 114 is to resolve conflicts between two or more modules which simultaneously wish to steer the harvester.
- the arbiter also performs a smoothing or filtering function on the steering direction votes produced by the various modules.
- This function may explained with reference to Fig. 11 which illustrates a typical situation for the case of a poor crop. The figure has been drawn for the case of less than 201 steering directions to more clearly demonstrate the smoothing function.
- graphs (a)-(d) represent the vote counts produced by the CLT module 102 for four successive images, graph (a) being for the most recent image.
- Each vertical bar in a graph represents the vote count in one of the bins, that is, the accumulated vote count for one bin as developed by the CLT module 102 at step 127 in Fig. 8 and before the vote counts are normalized.
- the vote for each image have been time weighted by a factor as indicated to the left of each graph.
- the bin or steering direction referenced 220 has the highest vote sum hence if the votes derived from this image alone are considered, the harvester would be commanded to make a rather extreme turn to the left because this bin is separated from the "straight ahead" bin, represented by 0, by about 15 bins.
- the vote summations for the bins referenced 221, 222 and 223 are the greatest hence if the vote summations for these images were acted on alone, the harvester would be commanded to steer gently to the right by slightly varying degrees.
- the Steering Arbiter module 114 time weights or decays each of the 201 normalized bin vote counts for the last n images processed by crop line tracker 102.
- Step 251 averages the decayed normalized bin vote counts obtained at step 250 to generate a steering preference signal for the crop line tracker. From the foregoing description it is evident that the steering preference signal comprises 201 bins or values, each value being between 0.0 and 1.0 and the total of all values being 1.0. The crop line tracker steering preference signal is then saved at step 252.
- the arbiter time decays and averages the bin vote counts produced by modules other than the CLT module 102 to develop steering preference signals for these modules.
- the time rate of decay of the bin vote counts from these modules need not be the same as for the bin vote counts from the CLT module.
- the bin vote counts may be averaged over more or fewer images than used in averaging the crop line tracker vote counts. The time rate of decay and the number of images taken into account in the averaging processed is dependent on the importance of a module to the overall steering operation and the frequency at which the module can process images.
- the time decaying of the bin vote summations (step 250) and the summing of the decayed bin vote summations for several images (step 251) provides an advantage which is not readily apparent from the foregoing description.
- the various modules generate bin vote summations at different rates, for example 2Hz or 5Hz, and steering commands are generated by SA module 114 at a 10Hz rate. This would result in quick changes in the steering command if the bin vote summations for prior images was not taken into account.
- step 254 checks a timer to see if .1 second has elapsed. If step 254 determines that .1 second has not elapsed since generation of the last steering command, steps 250-253 are repeated. After .1 second the test at step 254 proves true and the SA module produces a steering command as subsequently described.
- SA module 114 one purpose of SA module 114 is to resolve conflicts when two or more modules simultaneously wish to control the same harvester function.
- the arbiter in this case generates a steering command which is a compromise between the conflicting desires of the two modules.
- the steering preference signals for each module are weighted according to the importance of the module by multiplying each steering preference signal by a constant value. For example, the value in each bin of the crop line tracker steering preference signal is multiplied by a constant k. at 256 and the value in each bin of the steering preference signal for module M is multiplied by a constant k 2 at 257. The values in corresponding bins of the weighted preference signals are then added together at step 258. Step 259 examines each of the 201 bin totals obtained at step 258 and determines which bin has the largest total. The arbiter then generates (step 260) a steering command having a value which depends on the position of the bin found to have the largest total.
- the harvester transmission 50 is capable of steering in one of 5 directions.
- the crop line tracker and module M (either the LTT module 106, GTT module 106, or GTT module 108) steering preference signals, after weighting at 256 and 257, have the values 0.6, 0.4, 0, 0,0 and 0.1, 0.6, 0.1, 0.1, 0, respectively.
- the steering preference signal for the crop line tracker indicates a preference for a hard left steer (0.6) but a slight left turn (0.4) is acceptable.
- the steering preference signal for module B indicates a preference for a slight left turn (0.6).
- the bin totals are 0.7, 1.0, 0.1, 0.1, 0.1.
- Step 259 determines that the second bin from the left has the highest total (1.0) so at step 260 a steering command is generated to cause a slight turn to the left.
- the steering direction votes produced by the Obstacle Avoider module for each image contain veto votes (-1) in any bin associated with a steering direction which would steer the harvester on a collision course with a detected obstacle.
- these veto votes may be used in different ways to inhibit the generation of a command to steer along the collision course.
- the veto votes may inhibit the summing of values in those bin positions corresponding to the bin positions containing veto votes, or at step 259 the veto votes may exclude from the determination of the bin with the highest vote total those bins corresponding to the bins having veto votes therein.
- the steering commands produced by the Steering Arbiter module at step 260 are transferred from the Task Management Computer 52 to the Physical Control Computer 50.
- PCC 50 resolves the commands and generates analog signals which control valves in transmission 35 so that the wheel drive motors are differentially driven to steer the harvester in the commanded direction.
- a steering command is generated once each .1 sec when the test at step 254 (Fig. 12) proves true. After each steering command is generated the arbiter returns to step 250.
- the present invention provides a robotic control system capable of replacing an operator in controlling all functions of a harvester.
- the control program does not have to be changed in order to accommodate fields of different sizes or shapes because the path of harvester travel and other functions are specified in a field coverage plan that is planned within the system.
- An alternative embodiment provides a cruise control which relieves the operator of the responsibility of steering the harvester along a crop line but leaves all other functions under operator control.
Landscapes
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Environmental Sciences (AREA)
- Mechanical Engineering (AREA)
- Soil Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Guiding Agricultural Machines (AREA)
Claims (18)
- Système de contrôle robotique pour une récolteuse agricole (10), comprenant au moins une caméra vidéo (42/44) montée sur la récolteuse (10) afin de visualiser le changement de scène à l'avant de la récolteuse (10) et de produire des images du changement de scène lorsque la récolteuse (10) se déplace le long d'un trajet spécifié, et un module (104) destiné à balayer chaque ligne d'une image pour déterminer si une limite de ligne de récolte est indiquée dans la ligne,
caractérisé en ce que ledit module comprend un module de détecteur d'extrémité de bande (104) pour déterminer la ligne la plus probable représentant l'extrémité d'une bande de récolte à partir de la détermination des lignes indiquant une limite de ligne de récolte et des lignes ne l'indiquant pas. - Système de contrôle robotique selon la revendication 1, caractérisé en ce que ledit module de détecteur d'extrémité de bande (104) calcule une distance calculée à partir d'une position actuelle de récolteuse vers une position déterminée, correspondant à une position d'extrémité de bande la plus probable, ladite distance calculée étant comparée à une valeur de distance sauvegardée, ledit module de détecteur d'extrémité de bande (104) traitant une nouvelle image et déterminant une nouvelle position déterminée à partir de laquelle une nouvelle distance calculée est calculée, lorsque ladite distance calculée est supérieure à ladite distance sauvegardée.
- Système de contrôle robotique selon la revendication 2, caractérisé en ce que ledit module de détecteur d'extrémité de bande (104) continue à traiter des nouvelles images et à déterminer des nouvelles positions déterminées jusqu'à ce que la nouvelle distance calculée soit inférieure ou égale à ladite distance sauvegardée pour une nombre prédéterminé de positions calculées nouvellement déterminées.
- Système de contrôle robotique selon l'une quelconque des revendications précédentes, caractérisé en ce qu'il comporte en outre:un module de surveillance d'exécution (100) séquencé à travers une pluralité d'états finis par des messages de déclenchement issus d'un module de planificateur de couverture de champ (118);un module de suiveur de trajectoire globale (108);un module de contrôleur (120); etun module de surveillance de couverture de champ (118),
- Système de contrôle robotique selon la revendication 4, caractérisé en ce que:ledit module de suiveur de trajectoire globale (108) génère des propositions indiquant la direction préférée de conduite de la récolteuse (10) de manière à suivre ledit trajet; etledit module de suiveur de ligne de récolte (102) génère des propositions de conduite sur la direction de conduite de la récolteuse (10) pour suivre une ligne de récolte (16) entre le matériau de récolte fauché et le matériau de récolte non fauché,le système comprenant en outre un module d'arbitrage de conduite en fonctionnement continu (114) afin de développer des commandes de conduite à partir de toutes les propositions de conduite et un module de contrôleur en fonctionnement continu (120) réagissant aux commandes de conduite afin de développer des signaux pour entraíner différentiellement les roues avant (22, 24) de la récolteuse (10) afin de la diriger.
- Système de contrôle robotique selon la revendication 4 ou 5, caractérisé en ce que ledit module de suiveur de ligne de récolte (102) génère, pour chaque image, des propositions de conduite sur la direction de conduite de la récolteuse (10) afin de suivre une ligne de récolte (16) entre le matériau de récolte fauché et le matériau de récolte non fauché,
le système comportant en outre un module d'arbitrage de conduite fonctionnant en continu (114) afin de développer une commande de conduite à partir de toutes les propositions de conduite, et un module de contrôleur (120) en fonctionnement continu réagissant aux commandes de conduite afin de développer des signaux pour entraíner différentiellement les roues avant (22, 24) de la récolteuse (10) afin de la diriger. - Système de contrôle robotique selon la revendication 6, caractérisé en ce que, en développant les commandes de conduite, les propositions issues du module de suiveur de ligne de récolte (102) sont pondérées différemment des propositions issues du module de suiveur de trajectoire globale (108).
- Système de contrôle robotique selon l'une quelconque des revendications 4 à 7, caractérisé en ce que ledit module de suiveur de ligne de récolte (102) compense les images quant aux ombres dans la scène et dérive, de chaque image compensée, des propositions de conduite pour conduire la récolteuse (10) le long d'une ligne de récolte (16) entre le matériau de récolte fauché et le matériau de récolte non fauché.
- Système de contrôle robotique selon la revendication 8, caractérisé en ce que lesdites propositions de conduite sont amorties chronologiquement sur X images pour produire un signal préférentiel de conduite.
- Système de contrôle robotique selon l'une quelconque des revendications 4 à 9, caractérisé en ce que ledit module de suiveur de ligne de récolte (102) analyse chacune des images produites selon une fonction de discriminant de couleurs étant adaptativement mise à jour pour dériver des propositions de conduite afin de conduire la récolteuse (10) le long d'une ligne de récolte (16) entre le matériau de récolte fauché et le matériau de récolte non fauché.
- Système de contrôle robotique selon la revendication 10, caractérisé en ce que ledit module de suiveur de ligne de récolte (102) comporte:un discriminateur pour analyser chaque pixel de chaque ligne de chaque image produite et assigner une valeur de discriminant à chaque dit pixel; etun segmenteur pour calculer une meilleure corrélation de fonction discontinue pour les pixels de chaque ligne et définissant un pixel sélectionné à chaque ligne pour représenter l'emplacement de la ligne de récolte.
- Système de contrôle robotique selon la revendication 10 ou 11, caractérisé en ce que, après chaque analyse d'image, le module de suiveur de ligne de récolte (102) calcule le discriminant linéaire de Fisher dans le système RVB (rouge-vert-bleu) entre le matériau de récolte fauché et le matériau de récolte non fauché en tant que discriminant pour l'image suivante.
- Méthode de récolte de plantes avec une récolteuse (10) possédant un système de contrôle robotique, comprenant les étapes:de production d'images à partir d'au moins une caméra vidéo (42/44) montée sur ladite récolteuse (10) pouvant indiquer le changement de scène à l'avant de ladite récolteuse (10) lorsque ladite récolteuse se déplace le long d'un trajet de conduite; etde balayage de chaque ligne de chaque image à l'aide d'un module (104) afin de déterminer si une limite de ligne de récolte est indiquée;de détermination, par ledit module de détecteur de bande, de la ligne la plus probable représentant l'extrémité d'une bande de récolte à partir de la détermination des lignes indiquant une limite de ligne de récolte et des lignes ne l'indiquant pas.
- Méthode selon la revendication 13, caractérisée en ce qu'elle comporte en outre les étapes suivantes:de prévision d'un module de surveillance d'exécution (100) séquencé à travers une pluralité d'états finis par des messages de déclenchements issus d'un module de planificateur de couverture de champ (118), un module de suiveur de trajectoire globale (108), un module de contrôleur (120), ledit module de détecteur d'extrémité de bande (104), un module de suiveur de ligne de récolte (102) et un module de surveillance de couverture de champ (118); etd'activation d'au moins l'un des modules de planificateur de couverture de champ (118), de suiveur de trajectoire globale (108), de détecteur d'extrémité de bande (104), ou de suiveur de trajectoire locale (106) dans chaque état fini de manière à générer des signaux de conduite de la récolteuse afin de diriger la récolteuse le long d'un trajet spécifié par un plan de couverture de champ pour faucher tout le matériau de récolte dans un champ.
- Méthode selon la revendication 14, caractérisée en ce qu'elle comporte en outre les étapes:de production de propositions de conduite à partir dudit au moins module de trajectoire global, indiquant la direction préférée de la récolteuse (10) pour suivre ledit trajet, et ledit module de suiveur de ligne de récolte (102) définissant une ligne de récolte entre le matériau de récolte fauché et le matériau de récolte non fauché le long de laquelle la récolteuse (10) doit être conduite;d'arbitrage desdites propositions de conduite par un module d'arbitrage de conduite fonctionnant en continu (114) pour développer des commandes de conduite à partir de toutes les propositions de conduite et un contrôleur fonctionnant en continu (120) réagissant aux commandes de conduite pour développer des signaux afin d'entraíner différentiellement les roues avant (22, 24) de la récolteuse (10) pour la diriger; etde conduite de ladite récolteuse (10) en réponse auxdites commandes de conduite.
- Méthode de récolte de matériau de récolte, selon la revendication 15, caractérisée en ce qu'elle comporte en outre les étapes:de pondération desdites propositions de conduite à ladite étape d'arbitrage de sorte que lesdites propositions de conduite dudit module de suiveur de ligne de récolte (102) sont pondérées autrement que les propositions de conduite dudit module de suiveur de trajectoire globale (108).
- Méthode selon la revendication 16, caractérisée en ce qu'elle comporte en outre les étapes:de compensation des images quant aux ombres pour créer des images compensées; etde dérivation hors desdistes images compensées pour obtenir lesdites propositions de conduite pour ledit module de suiveur de ligne de récolte (102) afin de conduire la récolteuse (10) le long d'une ligne de récolte entre le matériau de récolte fauché et le matériau de récolte non fauché.
- Méthode selon la revendication 17, caractérisée en ce que ladite étape de dérivation comporte l'étape:d'analyse de chacune desdites images conformément à une fonction de discriminant de couleurs étant adaptativement mise à jour pour définir ladite ligne de récolte entre le matériau de récolte fauché et le matériau de récolte non fauché.
Applications Claiming Priority (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US4200397P | 1997-04-16 | 1997-04-16 | |
US42003P | 1997-04-16 | ||
US7916098P | 1998-03-24 | 1998-03-24 | |
US79160P | 1998-03-24 | ||
PCT/US1998/007314 WO1998046065A1 (fr) | 1997-04-16 | 1998-04-15 | Machine agricole de recolte commandee par robot |
Publications (3)
Publication Number | Publication Date |
---|---|
EP0975209A1 EP0975209A1 (fr) | 2000-02-02 |
EP0975209A4 EP0975209A4 (fr) | 2000-05-31 |
EP0975209B1 true EP0975209B1 (fr) | 2003-05-21 |
Family
ID=26718779
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP98918120A Expired - Lifetime EP0975209B1 (fr) | 1997-04-16 | 1998-04-15 | Machine agricole de recolte commandee par robot |
Country Status (5)
Country | Link |
---|---|
US (1) | US6336051B1 (fr) |
EP (1) | EP0975209B1 (fr) |
AU (1) | AU7110698A (fr) |
DE (1) | DE69814847T2 (fr) |
WO (1) | WO1998046065A1 (fr) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102436225A (zh) * | 2011-09-15 | 2012-05-02 | 西北农林科技大学 | 一种果园机械导航*** |
CN103914071A (zh) * | 2014-04-02 | 2014-07-09 | 中国农业大学 | 一种用于谷物联合收割机的视觉导航路径识别*** |
CN103914070A (zh) * | 2014-04-02 | 2014-07-09 | 中国农业大学 | 一种用于谷物联合收割机的视觉导航辅助驾驶*** |
Families Citing this family (163)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
AUPP679598A0 (en) | 1998-10-27 | 1998-11-19 | Agsystems Pty Ltd | A vehicle navigation apparatus |
FR2800232B1 (fr) * | 1999-10-27 | 2002-06-07 | Renault Agriculture | Procede pour le guidage automatique d'un engin et dispositif correspondant |
EP1278409B1 (fr) | 2000-05-03 | 2004-09-29 | Eco-Dan A/S | Procede de segmentation d'une image numerique, et procede de commande d'un instrument |
US6714662B1 (en) * | 2000-07-10 | 2004-03-30 | Case Corporation | Method and apparatus for determining the quality of an image of an agricultural field using a plurality of fuzzy logic input membership functions |
GB2386971B (en) * | 2002-03-26 | 2005-11-30 | Mcmurtry Ltd | Method of operating an automated land maintenance vehicle |
NL1020804C2 (nl) * | 2002-06-06 | 2003-12-09 | Lely Entpr Ag | Werkwijze en systeem voor het uitvoeren van ten minste twee landbouwbewerkingen op een landbouwperceel. |
NL1020789C2 (nl) * | 2002-06-06 | 2003-12-09 | Lely Entpr Ag | Werkwijze en inrichting voor het bepalen van de positie van een autonome landbouwmachine in een landbouwperceel. |
NL1020792C2 (nl) * | 2002-06-06 | 2003-12-09 | Lely Entpr Ag | Landbouwmachine voor het uitvoeren van een landbouwbewerking. |
NL1020798C2 (nl) * | 2002-06-06 | 2003-12-09 | Lely Entpr Ag | Landbouwmachine voor het uitvoeren van een landbouwbewerking, in het bijzonder een gewasbewerking. |
ES2208091B1 (es) * | 2002-07-24 | 2005-03-16 | Universidad De Malaga | Sistema robotizado para servicio en invernaderos. |
US6928339B2 (en) | 2002-11-21 | 2005-08-09 | Reinke Manufacturing Company, Inc. | GPS-based control system and method for controlling mechanized irrigation systems |
US8214111B2 (en) * | 2005-07-19 | 2012-07-03 | Hemisphere Gps Llc | Adaptive machine control system and method |
US9002565B2 (en) | 2003-03-20 | 2015-04-07 | Agjunction Llc | GNSS and optical guidance and machine control |
US8634993B2 (en) | 2003-03-20 | 2014-01-21 | Agjunction Llc | GNSS based control for dispensing material from vehicle |
US7010425B2 (en) * | 2003-03-31 | 2006-03-07 | Deere & Company | Path planner and a method for planning a path of a work vehicle |
US7228214B2 (en) * | 2003-03-31 | 2007-06-05 | Deere & Company | Path planner and method for planning a path plan having a spiral component |
US6907336B2 (en) * | 2003-03-31 | 2005-06-14 | Deere & Company | Method and system for efficiently traversing an area with a work vehicle |
US7216033B2 (en) * | 2003-03-31 | 2007-05-08 | Deere & Company | Path planner and method for planning a contour path of a vehicle |
US6934615B2 (en) * | 2003-03-31 | 2005-08-23 | Deere & Company | Method and system for determining an efficient vehicle path |
KR20040091788A (ko) * | 2003-04-22 | 2004-11-02 | 현대자동차주식회사 | 고속도로 자율주행 시스템 및 그의 제어방법 |
DE102004036946A1 (de) * | 2003-10-02 | 2005-05-19 | Claas Selbstfahrende Erntemaschinen Gmbh | Transformationsvorrichtung innerhalb einer landwirtschaftlichen Maschine |
US7079943B2 (en) * | 2003-10-07 | 2006-07-18 | Deere & Company | Point-to-point path planning |
US7593798B2 (en) * | 2003-10-30 | 2009-09-22 | Deere & Company | Vehicular guidance system having compensation for variations in ground elevation |
DE10351861A1 (de) | 2003-11-06 | 2005-06-09 | Deere & Company, Moline | Verfahren und Lenksystem zum selbsttätigen Lenken einer landwirtschaftlichen Maschine |
ATE536577T1 (de) | 2004-06-24 | 2011-12-15 | Irobot Corp | Fernbediente ablaufsteuerung und verfahren für eine autonome robotervorrichtung |
US11209833B2 (en) | 2004-07-07 | 2021-12-28 | Irobot Corporation | Celestial navigation system for an autonomous vehicle |
US8972052B2 (en) | 2004-07-07 | 2015-03-03 | Irobot Corporation | Celestial navigation system for an autonomous vehicle |
US7706917B1 (en) | 2004-07-07 | 2010-04-27 | Irobot Corporation | Celestial navigation system for an autonomous robot |
US7248968B2 (en) * | 2004-10-29 | 2007-07-24 | Deere & Company | Obstacle detection using stereo vision |
DE102004059543A1 (de) * | 2004-12-09 | 2006-06-29 | Claas Selbstfahrende Erntemaschinen Gmbh | Landwirtschaftliche Arbeitsmaschine |
US20060135232A1 (en) * | 2004-12-17 | 2006-06-22 | Daniel Willis | Method and system for delivering advertising content to video games based on game events and gamer activity |
US20060167600A1 (en) * | 2005-01-27 | 2006-07-27 | Raven Industries, Inc. | Architecturally partitioned automatic steering system and method |
DE102005005557C5 (de) | 2005-02-07 | 2019-03-14 | Pöttinger Landtechnik Gmbh | Vorrichtung zur Steuerung und/oder Regelung einer Landmaschine mittels einer Bilderfassungs- und -verarbeitungseinrichtung |
DE102005008105A1 (de) * | 2005-02-21 | 2006-08-31 | Amazonen-Werke H. Dreyer Gmbh & Co. Kg | Elektronisches Maschinen-Management-System |
CA2509888A1 (fr) * | 2005-06-14 | 2006-12-14 | Accutrak Systems Limited | Systeme et methode de pilotage automatique |
US7792622B2 (en) * | 2005-07-01 | 2010-09-07 | Deere & Company | Method and system for vehicular guidance using a crop image |
US7580549B2 (en) * | 2005-07-01 | 2009-08-25 | Deere & Company | Method and system for vehicular guidance using a crop image |
US7570783B2 (en) * | 2005-07-01 | 2009-08-04 | Deere & Company | Method and system for vehicular guidance using a crop image |
US8185275B2 (en) * | 2005-07-01 | 2012-05-22 | Deere & Company | System for vehicular guidance with respect to harvested crop |
US7684916B2 (en) * | 2005-07-01 | 2010-03-23 | Deere & Company | Method and system for vehicular guidance using a crop image |
US7707810B2 (en) * | 2005-07-15 | 2010-05-04 | Cnh America Llc | Apparatus and method to vary the reel speed versus ground speed of an agricultural windrower |
DE102005041550A1 (de) * | 2005-08-31 | 2007-03-01 | Agrocom Gmbh & Co. Agrarsysteme Kg | Lenksystem eines Fahrzeugs |
ATE454811T1 (de) * | 2005-09-14 | 2010-01-15 | Agrocom Verwaltungs Gmbh | Verfahren zur steuerung einer ballenpresse und ballenpresse |
US8944866B2 (en) | 2011-09-15 | 2015-02-03 | Liquid Robotics, Inc. | Wave-powered endurance extension module for unmanned underwater vehicles |
NZ592743A (en) | 2006-01-20 | 2011-09-30 | Liquid Robotics Inc | A wave-powered vehicle and fin system therefor |
US20070282812A1 (en) * | 2006-03-08 | 2007-12-06 | Superior Edge, Inc. | Process execution support system |
CA2539354A1 (fr) * | 2006-03-10 | 2007-09-10 | Straw Track Manufacturing Inc. | Guidage d'instruments aratoires |
US7369924B2 (en) * | 2006-06-05 | 2008-05-06 | Deere & Company | System and method for providing guidance towards a far-point position for a vehicle implementing a satellite-based guidance system |
US8648867B2 (en) | 2006-09-25 | 2014-02-11 | Neurala Llc | Graphic processor based accelerator system and method |
US7865285B2 (en) * | 2006-12-27 | 2011-01-04 | Caterpillar Inc | Machine control system and method |
US8060299B2 (en) | 2007-02-28 | 2011-11-15 | Caterpillar Inc. | Machine with automated steering system |
CN101622173B (zh) | 2007-03-02 | 2013-03-27 | 里奎德机器人技术公司 | 水上载体 |
DE102007056318A1 (de) * | 2007-04-12 | 2008-10-16 | Deere & Company, Moline | Kommunikationssystem eines Fahrzeugs und Verfahren zum Betreiben eines Kommunikationssystems |
US7739015B2 (en) * | 2007-07-31 | 2010-06-15 | Deere & Company | System and method for controlling a vehicle with a sequence of vehicle events |
US8209075B2 (en) | 2007-07-31 | 2012-06-26 | Deere & Company | Method and system for generating end turns |
US8635011B2 (en) | 2007-07-31 | 2014-01-21 | Deere & Company | System and method for controlling a vehicle in response to a particular boundary |
US20090088979A1 (en) * | 2007-09-27 | 2009-04-02 | Roger Dale Koch | Automated machine navigation system with obstacle detection |
US8915692B2 (en) * | 2008-02-21 | 2014-12-23 | Harvest Automation, Inc. | Adaptable container handling system |
US8131432B2 (en) | 2008-02-27 | 2012-03-06 | Deere & Company | Method and system for managing the turning of a vehicle |
US8204654B2 (en) * | 2008-03-20 | 2012-06-19 | Deere & Company | System and method for generation of an inner boundary of a work area |
CA2627053C (fr) * | 2008-03-26 | 2012-05-22 | Macdon Industries Ltd. | Tracteur avec direction automatique |
US7721830B2 (en) * | 2008-03-27 | 2010-05-25 | Macdon Industries Ltd. | Tractor with automatic steering arrangement |
US7793442B2 (en) * | 2008-04-29 | 2010-09-14 | Caterpillar Inc | Avoidance system for locating electric cables |
US7958982B2 (en) * | 2008-04-29 | 2011-06-14 | Caterpilar Inc. | Cable guide having a signaling instrument |
EP2135498B1 (fr) * | 2008-06-20 | 2012-06-20 | AGROCOM GmbH & Co. Agrarsystem KG | Procédé de navigation d'un véhicule agricole et véhicule agricole |
US8527155B2 (en) * | 2008-06-27 | 2013-09-03 | Caterpillar Inc. | Worksite avoidance system |
DE102008056557A1 (de) * | 2008-11-10 | 2010-05-12 | Claas Selbstfahrende Erntemaschinen Gmbh | Erstellung von Bilddatenbanken für Bildauswertung |
US8073609B2 (en) * | 2008-12-17 | 2011-12-06 | Caterpillar Inc. | Slippage condition response system |
US8140239B2 (en) * | 2008-12-17 | 2012-03-20 | Caterpillar Inc. | Slippage condition response system |
US8386129B2 (en) * | 2009-01-17 | 2013-02-26 | Hemipshere GPS, LLC | Raster-based contour swathing for guidance and variable-rate chemical application |
US20110022231A1 (en) * | 2009-07-25 | 2011-01-27 | Jeffrey Walker | Apparatuses, Systems and Methods for Automated Crop Picking |
US8332106B2 (en) * | 2009-10-21 | 2012-12-11 | Caterpillar Inc. | Tether tracking system and method for mobile machine |
US8224516B2 (en) * | 2009-12-17 | 2012-07-17 | Deere & Company | System and method for area coverage using sector decomposition |
US8635015B2 (en) | 2009-12-17 | 2014-01-21 | Deere & Company | Enhanced visual landmark for localization |
EP2368419B1 (fr) * | 2010-03-23 | 2016-08-10 | CLAAS E-Systems KGaA mbH & Co KG | Procédé pour détecter une structure dans un champ, procédé de direction d'un véhicule agricole et véhicule agricole |
US8994519B1 (en) * | 2010-07-10 | 2015-03-31 | William Fuchs | Method of controlling a vegetation removal system |
BE1019422A3 (nl) * | 2010-07-14 | 2012-07-03 | Cnh Belgium Nv | Werkwijze en toestel voor voorspellende sturing van een landbouwvoertuigsysteem. |
US8694454B2 (en) | 2011-02-17 | 2014-04-08 | Superior Edge, Inc. | Methods, apparatus and systems for generating, updating and executing a vegetation control plan |
US9524646B2 (en) | 2011-03-17 | 2016-12-20 | Liquid Robotics, Inc. | Navigation of a fleet of autonomous vessels in current and wind |
TWI557024B (zh) | 2011-03-17 | 2016-11-11 | 液體機器學股份有限公司 | 航海設備、其放置於水中的方法、其從水中移去的方法、其總成、其適用的浮體、其功能控制方法及獲取資訊的方法 |
EP3267275B1 (fr) | 2011-03-17 | 2019-10-16 | Liquid Robotics, Inc. | Dispositifs houlomoteurs conçus pour s'emboîter |
TWI522278B (zh) * | 2011-03-17 | 2016-02-21 | 液體機器學股份有限公司 | 船隊、自主波能物質分配船舶、水施肥系統、其中央控制單元及將肥料或物質分配到水的本體的限定區域中的方法 |
EP3208185B1 (fr) | 2011-06-28 | 2019-05-01 | Liquid Robotics, Inc. | Embarcation recueillant à la fois la poussée de locomotion et l'énergie électrique provenant du mouvement des vagues |
CN102323824B (zh) * | 2011-07-06 | 2013-10-16 | 李华 | 卫星导航农机自行走控制***及其控制方法 |
US9147173B2 (en) | 2011-10-31 | 2015-09-29 | Harvest Automation, Inc. | Methods and systems for automated transportation of items between variable endpoints |
US20130173321A1 (en) * | 2011-12-30 | 2013-07-04 | Jerome Dale Johnson | Methods, apparatus and systems for generating, updating and executing a crop-harvesting plan |
US8504256B2 (en) * | 2012-01-06 | 2013-08-06 | Cnh America Llc | Harvester cruise control |
US8937410B2 (en) | 2012-01-17 | 2015-01-20 | Harvest Automation, Inc. | Emergency stop method and system for autonomous mobile robots |
GB2511714B (en) * | 2012-02-10 | 2017-12-27 | Deere & Co | Method of material handling using one or more imaging devices on the transferring vehicle and on the receiving vehicle to control the material distribution |
US9030549B2 (en) | 2012-03-07 | 2015-05-12 | Blue River Technology, Inc. | Method and apparatus for automated plant necrosis |
US9288938B2 (en) | 2012-06-01 | 2016-03-22 | Rowbot Systems Llc | Robotic platform and method for performing multiple functions in agricultural systems |
US9113590B2 (en) | 2012-08-06 | 2015-08-25 | Superior Edge, Inc. | Methods, apparatus, and systems for determining in-season crop status in an agricultural crop and alerting users |
US20150379721A1 (en) * | 2012-09-28 | 2015-12-31 | Justin Bak | Windrow relative yield determination through stereo imaging |
US9241442B2 (en) | 2012-10-23 | 2016-01-26 | Daniel A. DIAZDELCASTILLO | Autonomous and remote control all purpose machine (ARCAPM) |
WO2014105927A1 (fr) * | 2012-12-28 | 2014-07-03 | Agco Corporation | Utilisation d'une rampe virtuelle pour diriger un équipement agricole le long du bord de zones travaillées/non travaillées |
WO2014105928A1 (fr) * | 2012-12-28 | 2014-07-03 | Agco Corporation | Procédé et applications de système de coordonnées locales sur la base d'un flux de données optiques au moyen de caméras vidéo |
US10390497B2 (en) | 2013-03-07 | 2019-08-27 | Blue River Technology, Inc. | System and method for plant treatment |
US9717171B2 (en) | 2013-03-07 | 2017-08-01 | Blue River Technology Inc. | System and method for automated odometry calibration for precision agriculture systems |
US10327393B2 (en) | 2013-03-07 | 2019-06-25 | Blue River Technology Inc. | Modular precision agriculture system |
USD738386S1 (en) * | 2013-03-29 | 2015-09-08 | Deere & Company | Display screen with an animated graphical user interface |
USD732576S1 (en) * | 2013-03-29 | 2015-06-23 | Deere & Company | Display screen or portion thereof with icon |
EP3000030A4 (fr) | 2013-05-22 | 2017-07-05 | Neurala Inc. | Procédés et appareil pour une architecture de temps d'exécution distribuée non spécifique itérative et son application à une intelligence en nuage |
EP2999940A4 (fr) * | 2013-05-22 | 2017-11-15 | Neurala Inc. | Procédés et appareil pour intégration sensorielle précoce et acquisition robuste de connaissances du monde réel |
US9097520B2 (en) | 2013-06-12 | 2015-08-04 | Caterpillar Inc. | System and method for mapping a raised contour |
WO2015006600A1 (fr) * | 2013-07-10 | 2015-01-15 | Agco Corporation | Distribution automatisée du travail dans un champ |
WO2015006675A2 (fr) | 2013-07-11 | 2015-01-15 | Blue River Technology, Inc. | Procédé de mesure et de sélection automatique de phénotypes |
US9392743B2 (en) | 2013-08-14 | 2016-07-19 | Rowbot Systems Llc | Agricultural autonomous vehicle platform with articulated base |
BR112016011577B1 (pt) | 2013-11-20 | 2021-01-12 | Rowbot Systems Llc | plataforma de veículo autônomo, sistema de plataforma de veículo autônomo, robô agrícola e método para a navegação autônoma de um robô agrícola |
US9626566B2 (en) | 2014-03-19 | 2017-04-18 | Neurala, Inc. | Methods and apparatus for autonomous robotic control |
CA2941250A1 (fr) | 2014-03-19 | 2015-09-24 | Neurala, Inc. | Procedes et appareil de commande robotique autonome |
US9489576B2 (en) | 2014-03-26 | 2016-11-08 | F12 Solutions, LLC. | Crop stand analysis |
US9841768B2 (en) * | 2014-03-28 | 2017-12-12 | Yanmar Co., Ltd. | Autonomous travelling service vehicle |
US9317785B1 (en) * | 2014-04-21 | 2016-04-19 | Video Mining Corporation | Method and system for determining ethnicity category of facial images based on multi-level primary and auxiliary classifiers |
BR102014012786A8 (pt) * | 2014-05-27 | 2022-08-30 | Marino Bavaresco Isaac | Controlador eletrônico para plantadeiras que utiliza localização fornecida por gps |
DE102014113001A1 (de) * | 2014-09-10 | 2016-03-10 | Claas Selbstfahrende Erntemaschinen Gmbh | Verfahren zur Steuerung eines Überladeprozesses |
JP2016114984A (ja) * | 2014-12-11 | 2016-06-23 | 福田 敏男 | 作業車両 |
EP3264892B1 (fr) | 2015-03-06 | 2020-10-07 | Blue River Technology Inc. | Système modulaire d'agriculture de précision |
KR102301536B1 (ko) * | 2015-03-10 | 2021-09-14 | 삼성전자주식회사 | 고해상도 전자 현미경 이미지로부터 결정을 분석하는 방법 및 그 시스템 |
SE540519C2 (sv) * | 2015-08-26 | 2018-09-25 | Vaederstad Holding Ab | Förfarande för styrning av lantbruksredskap, samt lantbruksredskap |
DE102015115510A1 (de) | 2015-09-15 | 2017-03-16 | Claas Selbstfahrende Erntemaschinen Gmbh | Fahrerassistenzsystem für eine landwirtschaftliche Arbeitsmaschine |
CN105251563B (zh) * | 2015-09-22 | 2018-04-24 | 中国矿业大学 | 一种齿辊式破碎机的自动调速控制***及控制方法 |
US9795074B2 (en) | 2015-10-27 | 2017-10-24 | Cnh Industrial America Llc | Automatic swath generation device and methods |
US10080325B2 (en) | 2015-10-27 | 2018-09-25 | Cnh Industrial America Llc | Predictive overlap control model |
US10314224B2 (en) | 2016-03-30 | 2019-06-11 | Autonomous Solutions, Inc. | Multiple harvester planner |
WO2017177284A1 (fr) | 2016-04-15 | 2017-10-19 | University Of Southern Queensland | Procédés, systèmes et dispositifs se rapportant à la détection d'ombre pour l'identification d'objets en temps réel |
US9826673B1 (en) * | 2016-05-27 | 2017-11-28 | Cnh Industrial America Llc | Swath acquisition system for an agricultural vehicle |
US10143126B2 (en) | 2016-06-10 | 2018-12-04 | Cnh Industrial America Llc | Planning and control of autonomous agricultural operations |
US10073460B2 (en) * | 2016-06-10 | 2018-09-11 | Trimble Inc. | Providing auto-guidance of a mobile machine without requiring a graphical interface display |
US10251329B2 (en) | 2016-06-10 | 2019-04-09 | Cnh Industrial Canada, Ltd. | Planning and control of autonomous agricultural operations |
WO2017042789A1 (fr) * | 2016-07-20 | 2017-03-16 | Universidad Tecnológica De Panamá | Robot d'entretien de pelouse et de récolte |
US10051774B2 (en) | 2016-09-16 | 2018-08-21 | Cnh Industrial Canada, Ltd. | Common lift/lower ground entrance point |
DE102016118227A1 (de) * | 2016-09-27 | 2018-03-29 | Claas Selbstfahrende Erntemaschinen Gmbh | Bildanlysesystem für landwirtschaftliche Arbeitsmaschinen |
US10255670B1 (en) | 2017-01-08 | 2019-04-09 | Dolly Y. Wu PLLC | Image sensor and module for agricultural crop improvement |
US10721859B2 (en) | 2017-01-08 | 2020-07-28 | Dolly Y. Wu PLLC | Monitoring and control implement for crop improvement |
JP6945303B2 (ja) * | 2017-01-20 | 2021-10-06 | 株式会社クボタ | 作業車、及び、他車監視システム |
KR102539520B1 (ko) * | 2017-01-24 | 2023-06-05 | 가부시끼 가이샤 구보다 | 콤바인, 수확기, 및 자동 조타 시스템 |
DE102017204158A1 (de) | 2017-03-14 | 2018-09-20 | Robert Bosch Gmbh | Verfahren zum Erkennen mindestens eines Fortpflanzungsorgans einer Pflanze |
US10212400B2 (en) | 2017-05-01 | 2019-02-19 | Cnh Industrial America Llc | Systems of acquiring image data for an autonomous work vehicle |
IT201700058505A1 (it) | 2017-05-30 | 2018-11-30 | Volta Robots S R L | Metodo di controllo di un mezzo di lavorazione del terreno sulla base di elaborazioni d'immagini e relativo sistema |
DE102017122711A1 (de) * | 2017-09-29 | 2019-04-04 | Claas E-Systems Kgaa Mbh & Co. Kg | Verfahren für den Betrieb einer selbstfahrenden landwirtschaftlichen Arbeitsmaschine |
US11076531B2 (en) | 2018-01-23 | 2021-08-03 | Deere & Company | State machine for multiple input-multiple output harvester control |
US11277956B2 (en) * | 2018-07-26 | 2022-03-22 | Bear Flag Robotics, Inc. | Vehicle controllers for agricultural and industrial applications |
US11470760B2 (en) * | 2018-08-13 | 2022-10-18 | Raven Industries, Inc. | Comparative agricultural obstacle monitor and guidance system and method for same |
DE102018122683A1 (de) * | 2018-09-17 | 2020-03-19 | Claas Selbstfahrende Erntemaschinen Gmbh | Landwirtschaftliche Erntemaschine |
CN109814551A (zh) * | 2019-01-04 | 2019-05-28 | 丰疆智慧农业股份有限公司 | 谷物处理自动驾驶***、自动驾驶方法以及自动识别方法 |
CN111492831A (zh) * | 2019-01-11 | 2020-08-07 | 洪锦民 | 远程栽种植物的方法及***、农具、计算机程序产品 |
DE102019202299B4 (de) * | 2019-02-20 | 2020-12-31 | Zf Friedrichshafen Ag | Verfahren zur Online-Kalibrierung und Kalibriereinrichtung |
US11224154B2 (en) | 2019-03-19 | 2022-01-18 | Cnh Industrial America Llc | Headland turn planning for a work vehicle |
DE102019204251A1 (de) * | 2019-03-27 | 2020-10-01 | Zf Friedrichshafen Ag | Verfahren und System zur Linienerfassung auf einer landwirtschaftlichen Nutzfläche |
CN110209155B (zh) * | 2019-04-09 | 2024-07-16 | 丰疆智能科技股份有限公司 | 用于农割作业机器的图像采集装置及其处理方法 |
CN110286670A (zh) * | 2019-04-09 | 2019-09-27 | 丰疆智能科技股份有限公司 | 多台自动收割机的行驶路径规划***及其方法 |
CN110209154B (zh) * | 2019-04-09 | 2022-10-14 | 丰疆智能科技股份有限公司 | 自动收割机的残留收割路径规划***及其方法 |
US10820478B1 (en) * | 2019-07-30 | 2020-11-03 | Cnh Industrial America Llc | System and method for providing a visual indication of field surface conditions |
FI20195678A1 (fi) * | 2019-08-14 | 2021-02-15 | Ponsse Oyj | Menetelmä ja järjestelmä metsäkoneessa |
DE102019006720A1 (de) * | 2019-09-25 | 2021-03-25 | Friedrich-Alexander-Universität Erlangen - Nürnberg | Autonomer Betrieb eines Fahrzeugs |
US11607703B2 (en) * | 2020-03-10 | 2023-03-21 | Cnh Industrial America Llc | Sensor system and related method for adjusting the operating position of sensors of an agricultural sprayer |
DE102020213979A1 (de) | 2020-11-06 | 2022-05-12 | Friedrich-Alexander-Universität Erlangen-Nürnberg, Körperschaft des öffentlichen Rechts | Automatisches Führen eines Regelpunktes eines Fahrzeuggespannes entlang einer Soll-Trajektorie |
US11793099B2 (en) * | 2020-11-10 | 2023-10-24 | Deere & Company | Harvester implement steering control system to prevent over-running harvester head end during turn |
GB202020322D0 (en) * | 2020-12-22 | 2021-02-03 | Agco Int Gmbh | Automation of agricultural machine operation |
US20220386527A1 (en) * | 2021-06-08 | 2022-12-08 | Deere & Company | Row sense operations with context sensing |
SE2250085A1 (en) * | 2022-01-31 | 2023-08-01 | Husqvarna Ab | Improved operation for a robotic work tool system |
DE102022111020A1 (de) | 2022-05-04 | 2023-11-09 | Man Truck & Bus Se | Hydrostatischer Antrieb für ein Nutzfahrzeug und Verfahren zum Betreiben eines hydrostatischen Antriebs eines Nutzfahrzeugs |
US20240188496A1 (en) * | 2022-12-07 | 2024-06-13 | Deere & Company | System and method for adjusting roller in rotary header |
Family Cites Families (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
FR2240678B1 (fr) | 1973-08-14 | 1978-01-13 | Fahr Ag Maschf | |
DE2555283C3 (de) | 1975-12-09 | 1980-12-18 | Maschinenfabrik Fahr Ag Gottmadingen, 7702 Gottmadingen | Anordnung zur selbsttätigen Lenkung eines landwirtschaftlich genutzten Fahrzeuges |
US4077488A (en) * | 1976-06-15 | 1978-03-07 | Sperry Rand Corporation | Guidance assist system for agricultural machines |
US4482960A (en) | 1981-11-20 | 1984-11-13 | Diffracto Ltd. | Robot tractors |
US4532757A (en) | 1983-09-30 | 1985-08-06 | Martin Marietta Corporation | Robotic fruit harvester |
US4700301A (en) | 1983-11-02 | 1987-10-13 | Dyke Howard L | Method of automatically steering agricultural type vehicles |
US4726175A (en) | 1987-02-02 | 1988-02-23 | University Of Kentucky Research Foundation | Harvester guidance control system |
JPS6434202A (en) * | 1987-07-30 | 1989-02-03 | Kubota Ltd | Working wagon of automatic conduct type |
US5444965A (en) | 1990-09-24 | 1995-08-29 | Colens; Andre | Continuous and autonomous mowing system |
AU621803B3 (en) * | 1991-08-19 | 1992-02-06 | Steven James Rees | Row crop, row deviation sensing device |
US5327709A (en) | 1993-03-12 | 1994-07-12 | Ford New Holland, Inc. | Adjustable header flotation mechanism |
US5442552A (en) | 1993-03-16 | 1995-08-15 | The Regents Of The University Of California | Robotic cultivator |
US5528888A (en) * | 1993-12-27 | 1996-06-25 | Fuji Jukogyo Kabushiki Kaisha | Autonomous mowing vehicle and apparatus for detecting boundary of mowed field |
US5544474A (en) * | 1995-02-21 | 1996-08-13 | Finkelstein; Zvi | System for harvesting crop items and crop harvesting tools used therewith |
DE19508944A1 (de) | 1995-03-13 | 1996-09-19 | Claas Ohg | Selbstlenkvorrichtung |
EP0801885B1 (fr) * | 1996-04-19 | 2002-01-09 | Carnegie-Mellon University | Guidage par rapport à ligne de récolte basé sur la vision |
US5906647A (en) * | 1996-09-03 | 1999-05-25 | Chrysler Corporation | Vehicle mounted guidance antenna for automated durability road (ADR) facility |
US6070673A (en) * | 1996-11-22 | 2000-06-06 | Case Corporation | Location based tractor control |
US5809440A (en) * | 1997-02-27 | 1998-09-15 | Patchen, Inc. | Agricultural implement having multiple agents for mapping fields |
DE19719939A1 (de) * | 1997-05-13 | 1998-11-19 | Claas Ohg | Automatisch lenkbare Erntemaschine |
US5899950A (en) * | 1997-07-07 | 1999-05-04 | Case Corporation | Sequential command repeater system for off-road vehicles |
-
1998
- 1998-04-15 DE DE69814847T patent/DE69814847T2/de not_active Expired - Lifetime
- 1998-04-15 AU AU71106/98A patent/AU7110698A/en not_active Abandoned
- 1998-04-15 WO PCT/US1998/007314 patent/WO1998046065A1/fr active IP Right Grant
- 1998-04-15 EP EP98918120A patent/EP0975209B1/fr not_active Expired - Lifetime
-
2000
- 2000-08-28 US US09/648,741 patent/US6336051B1/en not_active Expired - Fee Related
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102436225A (zh) * | 2011-09-15 | 2012-05-02 | 西北农林科技大学 | 一种果园机械导航*** |
CN103914071A (zh) * | 2014-04-02 | 2014-07-09 | 中国农业大学 | 一种用于谷物联合收割机的视觉导航路径识别*** |
CN103914070A (zh) * | 2014-04-02 | 2014-07-09 | 中国农业大学 | 一种用于谷物联合收割机的视觉导航辅助驾驶*** |
Also Published As
Publication number | Publication date |
---|---|
US6336051B1 (en) | 2002-01-01 |
WO1998046065A1 (fr) | 1998-10-22 |
DE69814847T2 (de) | 2004-01-22 |
EP0975209A1 (fr) | 2000-02-02 |
DE69814847D1 (de) | 2003-06-26 |
AU7110698A (en) | 1998-11-11 |
EP0975209A4 (fr) | 2000-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP0975209B1 (fr) | Machine agricole de recolte commandee par robot | |
EP0801885B1 (fr) | Guidage par rapport à ligne de récolte basé sur la vision | |
KR102589076B1 (ko) | 수확기 | |
EP1738630B1 (fr) | Procédé et système de conduite véhiculaire pour la récolte moissonnée | |
US10129528B2 (en) | Control arrangement and method for controlling a position of a transfer device of a harvesting machine | |
Pilarski et al. | The demeter system for automated harvesting | |
US6819780B2 (en) | Method and apparatus for automatically steering a vehicle in an agricultural field using a plurality of fuzzy logic membership functions | |
EP1738631B1 (fr) | Procédé et système de conduite véhiculaire pour la récolte moissonnée | |
US7580549B2 (en) | Method and system for vehicular guidance using a crop image | |
EP3316218B1 (fr) | Dispositif de commande et procédé pour commander une position d'un dispositif de transfert d'une moissonneuse | |
JP7381402B2 (ja) | 自動走行システム | |
WO1996017279A1 (fr) | Systeme de guidage de vehicule | |
EP3315007B1 (fr) | Dispositif de commande et procédé pour commander une position d'un dispositif de transfert d'une moissonneuse | |
EP3315006A1 (fr) | Dispositif de commande et procédé pour commander une position d'un dispositif de transfert d'une moissonneuse | |
WO2023055383A1 (fr) | Système de suivi de rangée de véhicules | |
AU691051B2 (en) | Vehicle guidance system | |
US11864494B2 (en) | AI-optimized harvester configured to maximize yield and minimize impurities | |
JP2021007385A (ja) | 農作業機 | |
EP4095642B1 (fr) | Détection et manipulation d'irrégularités de matériau | |
US11849668B1 (en) | Autonomous vehicle navigation | |
WO2020262287A1 (fr) | Machine d'exploitation agricole, système de déplacement autonome, programme, support d'enregistrement dans lequel un programme est enregistré, et procédé | |
WO2022123889A1 (fr) | Véhicule de travail, système de détection d'état d'objet, procédé de détection d'état d'objet, programme de détection d'état d'objet, et support d'enregistrement dans lequel un programme de détection d'état d'objet est enregistré | |
US11981336B2 (en) | Vehicle row follow system | |
US20240202966A1 (en) | Crop row detection system, agricultural machine having a crop row detection system, and method of crop row detection | |
WO2022124001A1 (fr) | Machine de travail agricole, programme de commande de machine de travail agricole, support d'enregistrement sur lequel le programme de commande de machine de travail agricole est enregistré, et procédé de commande de machine de travail agricole |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
17P | Request for examination filed |
Effective date: 19991028 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): DE DK FR GB IT |
|
A4 | Supplementary search report drawn up and despatched |
Effective date: 20000417 |
|
AK | Designated contracting states |
Kind code of ref document: A4 Designated state(s): DE DK FR GB IT |
|
RIC1 | Information provided on ipc code assigned before grant |
Free format text: 7A 01B 69/00 A |
|
17Q | First examination report despatched |
Effective date: 20020128 |
|
GRAH | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOS IGRA |
|
GRAH | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOS IGRA |
|
GRAA | (expected) grant |
Free format text: ORIGINAL CODE: 0009210 |
|
AK | Designated contracting states |
Designated state(s): DE DK FR GB IT |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: IT Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT;WARNING: LAPSES OF ITALIAN PATENTS WITH EFFECTIVE DATE BEFORE 2007 MAY HAVE OCCURRED AT ANY TIME BEFORE 2007. THE CORRECT EFFECTIVE DATE MAY BE DIFFERENT FROM THE ONE RECORDED. Effective date: 20030521 |
|
REG | Reference to a national code |
Ref country code: GB Ref legal event code: FG4D |
|
REF | Corresponds to: |
Ref document number: 69814847 Country of ref document: DE Date of ref document: 20030626 Kind code of ref document: P |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: DK Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT Effective date: 20030821 |
|
ET | Fr: translation filed | ||
PLBE | No opposition filed within time limit |
Free format text: ORIGINAL CODE: 0009261 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT |
|
26N | No opposition filed |
Effective date: 20040224 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: GB Payment date: 20130405 Year of fee payment: 16 Ref country code: DE Payment date: 20130409 Year of fee payment: 16 |
|
PGFP | Annual fee paid to national office [announced via postgrant information from national office to epo] |
Ref country code: FR Payment date: 20130430 Year of fee payment: 16 |
|
REG | Reference to a national code |
Ref country code: DE Ref legal event code: R119 Ref document number: 69814847 Country of ref document: DE |
|
GBPC | Gb: european patent ceased through non-payment of renewal fee |
Effective date: 20140415 |
|
REG | Reference to a national code |
Ref country code: DE Ref legal event code: R119 Ref document number: 69814847 Country of ref document: DE Effective date: 20141101 |
|
REG | Reference to a national code |
Ref country code: FR Ref legal event code: ST Effective date: 20141231 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: DE Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20141101 Ref country code: GB Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20140415 |
|
PG25 | Lapsed in a contracting state [announced via postgrant information from national office to epo] |
Ref country code: FR Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES Effective date: 20140430 |